Available online at www.sciencedirect.

com

Powder Technology 186 (2008) 80 – 88

www.elsevier.com/locate/powtec

Wall surface effects on particle–wall friction factor

in upward gas–solid flows
R. Mabrouk ⁎, J. Chaouki, C. Guy
Department of Chemical Engineering, Ecole Polytechnique de Montreal, C.P. 6078, succ. Centre-Ville, Montreal, Quebec, Canada H3C 3A7
Received 27 April 2006; received in revised form 1 October 2007; accepted 2 November 2007
Available online 21 November 2007

Abstract

The variation of the particle–wall friction factor along the riser is investigated in an Internal Circulating Fluidized Bed (ICFB) riser 1 m in
length and 0.052 m in diameter. The results obtained are based on calculating the normal and the shear forces at the wall under dynamic conditions
rather than the static ones usually obtained in shear box experiments.
The strength of the method used resides in the measurement technique applied to measure the particle velocity field in the riser. The radioactive
particle tracking program was developed for coaxial systems and is used to build dynamic pictures of particle trajectories in the vicinity of the wall
of the ICFB riser.
The experiments were conducted using sand (dp = 250 µm) and alumina (dp = 170 µm) materials in the gas velocity range between 2 and 12 m/s.
The most common correlations for calculating the particle–wall friction factor are reviewed and compared to the results obtained in this work. The
data obtained demonstrates that the particle–wall friction factor is not a constant value but changes along the riser and with change in the gas
superficial velocity. The results also show the effect of the roughness of the wall surface and define the particle–wall friction factor area.
© 2007 Elsevier B.V. All rights reserved.

Keywords: Particle–wall friction; Internal Circulating Fluidized Bed; Friction factor

1. Introduction correlation based on the assumption that there is neither

acceleration nor wall effects. Others believe in its importance;
Understanding the pressure changes of a system is very Yang [2], for instance, considered that the total pressure drop
important, particularly so for circulating systems and those consisted of three individual contributions: acceleration, gravity
operating in a fast fluidization regime. The importance of and both gas and particle–wall frictions. In 1998, Rautiainen
pressure drop was taken into consideration from the outset of
research in this field. It is considered one of the main variables,
the fastest and easiest parameter to diagnose the behavior of Table 1
Correlations for shear expression
gas–solid systems. The acceleration of particles, solid retention
and friction against the wall significantly contribute to the Expressions Authors
riser's total pressure drop and the pressure loop in circulating fs ð1  eÞqs u2s [12–15]
systems. spw ¼
2
The importance of the pressure drop due to particle–wall [2,16,17]
fs ð1  eÞqs u2s
friction was, and still is, the subject of many debates. Some spw ¼
8
authors simply ignore its effect and consider it negligible. For
fs qf u2g [18,19]
example, Yerushalmi and Cankurt [1] developed a pressure drop spw ¼
8

fs qf u2g Ws [20]
⁎ Corresponding author. Fax: +1 514 340 4159. spw ¼
2 Wg
E-mail address: rachid.mabrouk@polymtl.ca (R. Mabrouk).

0032-5910/$ - see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.powtec.2007.11.009
 R. Mabrouk et al. / Powder Technology 186 (2008) 80–88 81

Table 2
Particle–wall friction factor correlations reported in the literature
Authors Correlation Column [m] Particle property
D H dp [µm] ρs [kg/m3]
[18]


 .126–.025 2.44 26–392 893.7–3118
Wg Ut D
fs ¼ 0:5 ð1  eÞ
Ws Ug dp
[10] fs = 0.003 .051 10 65 1600
[11] fs = 0.0015–0.003 .038–.050 6 20–290 868–2740
[13] fs = 0.048u−s 1.22 .0762 4.87 470–3.4E3 911–7850
[2] (ug/ut〉1.5)   –
ð1  eÞ:021 Rep :979 .068–.076 120–3.25E3 910–8900
fs ¼ 0:0126
e3 Rept
(ug/ut〈1.5)  
ð1  eÞ:021 Rep 1:021
fs ¼ 0:041 :3
e Rept
[12] fs = 0.074u−s 0.75 .040 16.2 683–2.24E3 802–1154
[21] es .038 2 296–452 2180–2560
fs ¼ 12:2 3
e us


[22] es Ut es Ut 1:5 .030 1.2 1.2E3–2.9E3 2507–2641
fs ¼ 0:0017
e4 ug ug  us
[14]
 .192 4.4 64–310 2450
ub
fs ¼ fs;l 1  s
fs;l
[23] Ws .05 – 80–3E3 958–2650
fs ¼ 0:0108 þ 0:066
us Aqs

and Sarkomaa proposed a model in which the contribution of
the pressure drop due to gravity and friction against the wall
were counted.
Reviewing the literature there is strong evidence that the
flow field is greatly affected by many phenomena, such as
segregation and agglomeration in the vicinity of the wall. The
wall friction mechanism between particles and the wall,
expressed in terms of friction forces, influences the conveyance
of the solid and gas–solid mixing.
The pressure drop due to particle–wall friction is defined by
the expression DPpw ¼ 4spwDDZ , where τpw is particle–wall shear,
usually described by one of the expressions presented in Table 1.
The expressions were developed based on analogy to the
Fanning equation. Expression (1) reported in Table 1 is commonly
used to evaluate the contribution of particle–wall friction to total
pressure drop. The parameter fs defines the particle–wall friction
factor that is traditionally determined by:
• How does the particle–wall friction factor change with
changing velocity?
• How does the particle–wall friction factor vary with height?
• How does the state of the surface wall affect the particle–
wall friction factor?
The usual relationship between shear τ and normal stress σ is
given by the Coulomb failure condition [5,6], which is the
constitutive equation for a powder as represented in the follow-
ing equation:
Shear = Cohesion + Normal stress ⁎ Tangent of internal fric-
tion angle.
s ¼ kc þ r tan h: ð5Þ
Thus, from Eq. (5), the particle–wall friction coefficient fs
might be written as:

• using planar shear cell equipment [3,4]; sws kwc

fs ¼ ¼ þ tan h ð6Þ
• fitting the measured pressure drops. r r

A large number of correlations have been proposed in the where kwc defines particle–wall cohesiveness.
literature in order to predict the particle–wall friction factor, A similar definition was given by Amonton 300 years ago,
which can be defined as: a constant; a function of particle known as Amonton's law [7]. It states that the particle–wall
velocity; a function of dimensionless numbers; or a function of friction factor is inversely proportional to the normal forces
both solid holdup εs and particle velocity. Some of the exerted on the solid.
expressions for particle–wall friction factors found in the sws sws =S Fz
literature are summarized in Table 2. fs ¼ ¼ ¼ ; ð7Þ
r r=S Fn
In this work the particle–wall friction coefficient is measured
and investigated along the riser in order to clarify the questions where Fz and Fn are the tangential and normal components of
below. the forces exerted on the particle.
82 R. Mabrouk et al. / Powder Technology 186 (2008) 80–88

Knowing that:
 
PY d mp :Y
us
F ext: ¼ ;
dt
ðNewton0 sÞ
PY mp :d Y
us
F ext: ¼ ; mp is the particle weight ¼ const: ð8Þ
dt

Thus,

Fz m p  u iþ1
s;z  u i
s;z
fs ¼ ¼  ð9Þ
Fn mp  uiþ1  ui
s;r s;r

i i +1 i i +1
where us,r , us,r , us,z and us,z are the radial and axial
components of the particle velocities before and after the
collision (see Fig. 1).

2. Experimental setup and procedure

The experimental apparatus is illustrated in Fig. 2. It is a cold

model representing an ICFB. Briefly, the ICFB consists of two
coaxial tubes of different heights, which delimit two zones, a
riser zone and an annulus zone. The two zones are characterized
by two different regimes: a moving bed in the annulus and a fast
fluidization in the riser. The solid is fed into the annulus zone
where it is maintained at minimum fluidization through the
orifices made in the bottom of the riser. The solid travels from
the annulus to the riser through the orifices and is then conveyed
along the central tube. At the exit the solid is separated from the
gas via an impact device made for this purpose. The primary air
enters the unit from the bottom of the riser and the secondary air
enters laterally via the annulus zone. Primary and secondary airs
are completely independent.
Fig. 2. Experimental setup.
In order to minimize the electrostatic forces generated in both
the riser and the annulus, thin metallic wires, connected to the
ground, were implemented close to the wall of the ICFB.
Two sets of experiments were run for two extreme riser wall
surface states. The first set of experiments was run for a smooth
riser wall surface, while the second set of experiments was run
for an extremely rough riser wall surface. The rough surface was
obtained by covering the internal riser wall surface with
sandpaper.
Several experiments were run for different solids, at various
velocities ranging from 2 to 12 m/s. Air was used, under ambient
conditions, as a fluidizing medium. Table 3 summarizes some of
the physical properties of the solids used. The particle size
distribution of both alumina and sand particles is presented in
Fig. 3.
The behavior of the solid close the riser walls is captured by
the behavior of the tracer mixed with the solid circulating in the
ICFB unit. To insure the tracer follows the same path as the
solid circulating in the unit, mixtures of scandium oxide (the
isotope) and epoxy are made in such proportions that the density
and the diameter of the tracer are the same as the solid
circulating in the unit. Table 4 summarizes the tracer properties
used.

Table 3
Physical properties of the bed material
Material Diameter Density Minimum fluidization Terminal
(μm) (kg/m3) velocity (m/s) velocity (m/s)
Sand 250 2500 0.0622 1.80
Alumina 170 3400 0.0395 1.40
Fig. 1. Schema of particle–wall collision.
 R. Mabrouk et al. / Powder Technology 186 (2008) 80–88
Fig. 3. Particle size distribution (a) sand particles, (b) alumina particles.
Using radioactive particle tracking (RPT) technique we were
able to locate the tracer at each time step. From space
coordinates (x, y and z) and time step, particle velocity and
acceleration fields are build. Particle trajectories within 5 mm
thickness from the wall were selected for the particle–wall
friction factor calculations as it is explained above.
3. Measurement technique
A radioactive particle tracking (RPT) technique is used in
this work. A radioactive tracer tracks the behavior of the solid in
the ICFB. The photons emitted from the tracer are counted by
discriminators, which are linked to the detectors surrounding the
column (see Fig. 2). The counts registered by each of the
detectors are used to locate the tracer (x, y, z) at each time point,
with a relative error of 1.4%. A sampling time of 10 ms was
chosen for these experiments. A set of particle trajectories
spanning 5 h was produced for each experiment. From those
Table 4
Tracer properties
Tracer Density Diameter
(kg/m3) (μm)
Tracer A: used for the experiments with alumina particles 3340
Tracer B: used for experiments with sand particles 2540
83
trajectories and the sampling time, the dynamic properties of the
solid were calculated as follows:
 
up up;r ¼ dr=dt; up;h ¼ r  dh=dt; up;z ¼ dz=dt :
The program, developed in earlier years by Larachi et al.
[8,9], is used with some modifications to account for the
annulus and riser zones, which are characterized by two
different attenuations.
Using the measurement technique and the methodology
explained above, the particle–wall friction factor was calculated
along the riser by expression (7).
The particle–wall friction factor values reported here are
averaged values that were calculated for each of the five
segments forming the riser height. Samples of particle–wall
friction factor distributions in different segments of the riser
height are presented in Fig. 4. From the relative uncertainty
given earlier for the particle position (1.4%), particle–wall
friction factor uncertainty is estimated to be 11%.
4. Results and discussion
The results obtained are presented in three parts. The first
and second parts are answering the first two questions raised in
the Introduction and providing evidence that the particle–wall
friction factor is not a constant, while the third part discusses the
effect of the riser wall surface.
4.1. Variation of particle–wall friction factor versus riser
height
The variation of the particle–wall friction factor along the
riser height is investigated for the first time here. In Fig. 5, the
particle–wall friction factors are presented for sand particles
along the height of the riser at several gas superficial velocities.
The first observation to note in Fig. 5 is that for the five
velocities studied the same trend occurs. The particle–wall
friction factor changes along the riser from a high value close to
185
Fig. 4. Particle–wall friction factor distributions in different segments of the
260
riser.
84 R. Mabrouk et al. / Powder Technology 186 (2008) 80–88
Fig. 5. Particle–wall friction factor versus height of the riser (sand particles).
the entrance and decreases gradually to a lower and almost
constant value in the rest of the riser.
The same trend is observed and reported in Fig. 6 for alumina
particles, at different gas superficial velocities. Relatively high
particle–wall friction factor values occur at the bottom of the
riser and decrease gradually to almost constant values at the top
of the riser.
We believe this to be an effect of the riser entrance. The fact
that the solid enters laterally from the orifices at the lower part
of the riser (see Fig. 1) has an effect on the solid particle velocity
field, such that at the entrance, the radial component of the
velocity vector is more important than anywhere else in the
riser. In this particular case, the particle velocity field changes as
function of time and the intensity of the gas velocity field.
In other words, for a given gas superficial velocity the
particles move upward as time progresses, so that in the first
moments of solid–gas contact, the particle velocity field
changes as the particles are entrained in the direction of the
gas. As a result of such change the particle velocity components
(axial and radial) change. The radial component decreases while
the axial component increases, which affects the angle between
the particle and gas velocity vectors as time passes, i.e., as z
increases. As a consequence, in the first moments, when the
particle is still in the riser entrance region, the angle formed
Fig. 6. Particle–wall friction factor versus height of the riser (alumina particles).
between the particle and gas velocity vectors is bigger compared
to the one in the rest of the riser, where the particle velocity field
is almost parallel to the gas velocity field.
To eliminate any doubt concerning the magnitude of the
particle–wall friction factor at the riser entrance, two verifica-
tions were made. Friction factors were recalculated using data
with a higher sampling time (30 ms) to verify if other
phenomena are included when the sampling time is longer.
Fig. 7 shows clearly that there is no significant difference
between the experiments with the sampling times of 10 ms and
30 ms.
Higher particle–wall friction factor values at the entrance
part of the riser are noticed before they reach a constant in the
rest of the riser for both sets of experiments.
To complete our investigation into a possible effect of the
sampling time chosen, an attempt to model the particle
trajectory of a single spherical particle in a uniform gas flow
was made. Experimentally, it was impossible with the actual
acquisition system to run experiments for less than 5 ms.
In general, the particle trajectory is calculated using the
equation:
dYvp Y Y Y Y Y Y
qs Vp ¼ F D þ F g þ F b þ F Basset þ F Added þ F ele: ð10Þ
dt
In the following we assumed that the impact of the last three
forces (Basset, added mass and electrostatic force) was
negligible. The projection of expression (10) in a two-
dimensional space is given as follows:
The particle velocity components us,r and us,z versus time
were obtained by solving the differential equation system
described above using the Runge Kutta method.
Fig. 7. Particle–wall friction factor versus height of the riser at different
sampling times.
 R. Mabrouk et al. / Powder Technology 186 (2008) 80–88 85

 
The variation of the angle η, defined as Arctg us;r =us;z versus
time at different gas superficial velocities is presented in Fig. 8.
Two main conclusions might be gleaned from the simulation
results presented in Fig. 8, which supports what is explained
above. It demonstrates that, by increasing the gas velocity the
time needed for particles to have the gas direction decreases.
Fig. 8 also shows that for the sampling time chosen for the
experiments of this work (10 ms), the angle of the particle
vectors could not affect the values obtained for the particle–wall
friction factor at the riser entrance, but for smaller sampling
times and low gas superficial velocity, its effect could be
significant.
In the following paragraphs the particle–wall friction factor
average along the riser was considered.
Fig. 9. Particle–wall friction factor versus gas superficial velocity.
4.2. Variation of particle–wall friction factor versus gas
superficial velocity
and we can see that the total pressure drop decreases as gas
The variation of the particle–wall friction factor as a function superficial velocity increases. A result we believe is due to the
of the gas superficial velocity is presented in Fig. 9 for both sand presence of less solid in the riser as the velocity increases. In
and alumina particles. The figure shows a parabolic profile fact, the curves presented in Fig. 10(b) confirm our hypothesis
decreasing from high values at low gas superficial velocities to about the effect of the gas superficial velocity on solid
low values at high gas superficial velocities. At high gas concentration inside the ICFB riser. Fig. 10(b) clearly shows
superficial velocity, one may observe that, the particle–wall the decrease of solid hold-up as the gas superficial velocity
friction factor tends to a constant of about 0.005. increases.
Fig. 9 shows clearly that the particle–wall friction factor is Fig. 11 presents the variation of the particle–wall friction
not constant as was reported earlier in the literature by factor versus the total pressure drop in the ICFB riser. The
Stemerding [10] and Yousfi and Gau [11]. Our finding matches results obtained shows that the particle–wall friction factor
the results obtained by Kmiec et al. [12] and Capes and decreases quasi-linearly as the total pressure drop decreases.
Nakamura [13] in terms of variability of the particle–wall
friction factor. 4.4. Comparison of different results

4.3. Variation of particle–wall friction factor versus the total In order to compare the global results obtained with the
pressure drop results found in the literature, an average of the particle–wall
friction factors in the whole riser is calculated.
To better introduce the variation of the particle–wall friction As shown in Fig. 12, the particle–wall friction factor
factor versus the total pressure drop, in Fig. 10 are presented the decreases, while particle velocity increases. The same tendency
change of the solid hold-up and the total pressure drop inside the
ICFB riser versus the gas superficial velocity.
Fig. 10(a) presents the variation of the total pressure drop,
measured in the ICFB riser, versus the gas superficial velocity

Fig. 8. Variation of the angle between the particle and gas flow fields versus Fig. 10. Total pressure drop and solid hold-up in the ICFB riser versus gas
time. superficial velocity.
86 R. Mabrouk et al. / Powder Technology 186 (2008) 80–88

Fig. 13. Particle–wall friction factor versus gas superficial velocity (alumina
Fig. 11. Particle–wall friction factor versus total pressure drop in the ICFB riser.
particles).

is reported in the literature (Fig. 5), except for Rautiainen and

Sarkomaa [14] who proposed a correlation where the particle– rough riser wall surface. The same trend is conserved as in the
wall friction factor increases from lower values to higher and case of the smooth riser wall, but with higher values. The
constant values as the particle velocity increases. particle–wall friction factor increases with an increase in the
Fig. 12 also shows the differences between the predictions of roughness of the wall surface.
some expressions proposed in the literature, which might be
explained by small but significant differences [13,14,24,25,26]
in the operating conditions.
We believe that one of the main reasons for differences in
predictions proposed in the literature is due to the condition of
the riser wall. In almost all the works published in this subject
few studies reported the state of the wall. The results obtained
from the second set of experiments are presented below.

4.5. Variation of particle–wall friction factor versus roughness

of the wall surface

The results were obtained in the same system as the one

described above but the internal surface of the riser wall was
covered with sandpaper to simulate an extremely rough surface.
Fig. 13 shows the variation of the particle–wall friction
factor versus the gas superficial velocity in a system with a

Fig. 12. Particle–wall friction factor versus particle velocity in the riser (sand Fig. 14. (a) Particle–wall friction factor area versus particle velocity.
particles). (b) Particle–wall friction factor area versus (U–Umf).
 R. Mabrouk et al. / Powder Technology 186 (2008) 80–88 87

Depending on the state of the wall surface, a domain area is

defined by two curve expressions:

fss ¼ 0:022v1:0
s
¼ 0:031ðU  Umf Þ0:92 for a smooth wall surface; and

fsr ¼ 0:051v1:0
s
¼ 0:072ðU  Umf Þ0:92 for a rough wall surface:

The two curves delimit the possible position of the particle–

wall friction factor depending on gas or particle velocities.

Nomenclature
Ap Surface of spherical particle [m2]
Fig. 15. Particle–wall friction factor versus particle velocity in the riser (sand
CD Drag coefficient
particles). D Column diameter in [m].
L Column length in [m].
R Riser radius in [m].
qud
Since the roughness of the wall surface is a dynamic Rep Reynolds number Rep ¼ f As p
q Ud
mechanism, in other words this property changes with time and Rept Reynolds number Rept ¼ f At p
the number of operating hours, one cannot speculate about an U Gas superficial velocity [m/s].
exact value for the particle–wall friction factor. A boundary for Umf Minimum fluidization velocity [m/s].
this factor, however, can be determined. Ut Solid terminal velocity [m/s].
Fig. 14 presents the domain where the particle–wall friction Vp Volume of spherical particle [m3]
factor could be located. Ws Solid mass flow rate [kg/s].
The domain presented in Fig. 14(a) is delimited by a lower Wg Gas mass flow rate [kg/s].
curve defined by the expression dp Mean particle diameter [microns].
fs Particle–wall friction factor.
fss ¼ 0:022v1:0
s ð12Þ r Distance from the center of the riser in [m].
t Time in [s].
for particle–wall friction factors for a smooth wall surface, and
ug Gas velocity in [m/s].
an upper curve defined by the expression
vs Solid velocity in [m/s].
fsr 0:051v1:0
s ð13Þ
Greek letters
for particle–wall friction factors for an extremely rough wall ρf Fluid density in [kg/m3].
surface. ρs Particles density in [kg/m3].
Correlations (12) and (13) were presented as a function of ε Void fraction.
particle velocity, which is not always an accessible variable, so
corresponding expressions were presented below as a function
References
of gas superficial velocity.

fss ¼ :031ðU  Umf Þ0:92
fsr ¼ 0:072ðU  Umf Þ0:92 :
Another interesting result presented in Fig. 15 shows that
almost all the expressions found in the literature and reported in
Fig. 11 are located in the area limited by the two curves (the one
for a smooth surface and the other for a rough surface), which
confirms our suspicion about the effect of the state of surface
wall in predicting the particle–wall friction factor.
5. Conclusion
The experimental data reported in this article shows that the
particle–wall friction factor is not constant along the riser nor
when changing the gas superficial velocity.
ð14Þ [1] J. Yerushalmi, N.T. Cankurt, Further studies of the regimes of fluidization,
Powder Technol., 24 (2) (1979) 187–205.
[2] W. Yang, A correlation for solid friction factor in vertical pneumatic
ð15Þ conveying lines, AIChE J. 24 (1978) 548–552.
[3] A.W. Jenike, Quantitative design of mass-flow bins, Powder Technol. 1 (4)
(1967) 237–244.
[4] H. Abou-Chakra, U. Tuzun, Coefficient of friction of binary granular
mixtures in contact with a smooth wall. Part I: direct shear box
measurements of the effects of particle size ratio and particle surface
roughness, Chem. Eng. Sci. 54 (24) (1999) 5901–5912.
[5] V.V. Sokolovskii, Statistics of Granular Media, Pergamon Press, London
UK, 1965.
[6] M.B. Abbott, An Introduction to the Method of Characteristics, Elsevier,
New York, 1966.
[7] J.P.K. Seville, U. Tüzün, R. Clift, Processing of Particulate Solids, Blackie
Academic & Professional, London, 1997.
[8] F. Larachi, G. Kennedy, J. Chaouki, A [gamma]-ray detection system for 3-D
particle tracking in multiphase reactors, Nuc. Instrum. Methods Phys. Res.
Sect. A, Accel., Spectrom., Detect. Assoc. Equip. 338 (2–3) (1994) 568–576.
88 R. Mabrouk et al. / Powder Technology 186 (2008) 80–88
[9] F. Larachi, J. Chaouki, G. Kennedy, 3-D mapping of solids flow fields in
multiphase reactors with RPT, AIChE J. 41 (2) (1995) 439–443.
[10] S. Stemerding, Pneumatic transport of cracking catalyst in vertical risers,
Chem. Eng. Sci. 17 (1962) 599–608.
[11] Y. Yousfi, G. Gau, Aerodynamique de l'écoulement vertical de suspen-
sions concentrées gaz–solide 2: Chute de pression et vitesse relative gaz–
solide, Chem. Eng. Sc., 29 (9) (1974) 1947–1953.
[12] A. Kmiec, S. Mielczarski, J. Pajakowska, Experimental study on
hydrodynamics of a system in pneumatic flash dryer, Powder Technol.
20 (1) (1978) 67–74.
[13] C.E. Capes, K. Nakamura, Vertical pneumatic conveying: an experimental
study with particles in the intermediate and turbulent flow regimes, Can.
J. Chem. Eng. 51 (1973) 31–38.
[14] A. Rautiainen, P. Sarkomaa, Solids friction factors in upward, lean gas–
solids flows, Proceedings of the 1996 Conference on Pneumatic and
Hydraulic Conveying Systems, Apr 21–25, Powder Technol., vol. 95 (1),
1998, pp. 25–35.
[15] O.H. Hariu, M.C. Molstad, Pressure drop in vertical tubes in transport of
solids by gases, Ind. Eng. Chem. 41 (1949) 1148–1160.
[16] W. Yang, Estimating the solid particle velocity in vertical pneumatic
conveying lines, Ind. Eng. Chem. Fundam. 12 (1973) 349–352.
[17] C.Y. Wen, A.F. Galli, Dilute phase systems, in: J.F. Davidson, D. Harrison
(Eds.), Fluidization Proceeding, Academic press inc., New York, 1971.
[18] K.J.R. Sarma, M.R. Narasinga, Pneumatic conveyance and continuous
fluidization of solids, Ind. Eng. Chem. 51 (1959) 1449–1452.
[19] J.H. Jones, W.G. Braun, T.E. Daubert, H.D. Allendorf, Estimation of
pressure drop for vertical pneumatic transport of solids, AIChE J. 13
(1967) 608–611.
[20] W. Barth, Flow problems with mixtures of gases and entrained solid particles:
Stroemungstechnische probleme bei der foerderung von staubluft-
gemischen, Eng. Dig. 23 (11) (1962) 81–85.
[21] R.W. Breault, V.K. Mathur, High-velocity fluidized bed hydrodynamic
modeling. 1. Fundamental studies of pressure drop, Ind. Eng. Chem. Res.
28 (6) (1989) 684–688.
[22] R.V. Garic, Z.B. Grbavcic, S.Dj. Jovanovic, Hydrodynamic modeling of
vertical non-accelerating gas–solids flow, Powder Technol. 84 (1) (1995)
65–74.
[23] M. Lech, Mass flow rate measurement in vertical pneumatic conveying of
solid, Powder Technol. 114 (1–3) (2001) 55–58.
[24] H. Konno, S. Saito, Conveying of solids through horizontal and vertical
pipes, J. Chem. Eng. Jpn. 2 (1969) 211–217.
[25] K.V.S. Reddy, D.C.T. Pei, Particle dynamics in solids–gas flow in vertical
pipe, Ind. Eng. Chem. Fundam., 8 (3) (1969) 490–497.
[26] W.P.M. Van Swaaij, C. Buurman, J.W. van Bruegel, Shear stresses on the wall
of a dense gas–solids riser, Chem. Eng. Sci. 25 (11) (1970) 1818–1820.